Technical Documentation

pip install aga-governance

from aga import AgentSession

with AgentSession(gateway_id="my-gateway") as session:
    # Record each tool call decision
    session.record_tool_call(
        tool_name="read_file",
        decision="PERMITTED",
        reason="path within allowed prefix",
        request_id="req-001",
        arguments={"path": "/data/report.csv"},
    )
    session.record_tool_call(
        tool_name="write_file",
        decision="DENIED",
        reason="path outside allowed prefix",
        request_id="req-002",
        arguments={"path": "/etc/shadow"},
    )

    # Export and verify
    bundle = session.export_bundle()
    result = session.verify()
    print(f"Valid: {result['overall_valid']}")  # True

# CLI verification
python -m aga verify evidence-bundle.json

# Python API
from aga import verify_bundle_file
result = verify_bundle_file("evidence-bundle.json")
print(result["overall_valid"])  # True

# The MCP gateway sits between AI agents and tool providers.
# Every tool call is intercepted, evaluated against policy,
# and recorded as a signed governance receipt.

# Verify a gateway-produced evidence bundle:
curl -s https://aga-mcp-gateway.attestedintelligence.workers.dev/bundle \
  | python -m aga verify -